Phylogenetic tree reconstruction is a prominent problem in computational biology. Currently, all computational methods have their limitations and work well only for simple problems of small size. No existing method can guarantee that trees constructed for real-world problems are true phylogenetic trees for large and complex problems mainly because the existing computational models are not very biologically realistic. It has become a serious issue for many important real-life applications which often desire accurate results from phylogenetic analysis. Thus, it is very crucial to effectively incorporate multi-disciplinary analyses and synthesize results from various sources when answering real-life questions.
In this thesis, a novel web-based phylogeny reconstruction system with a real-time interactive environment, called Qphyl (short for quartet-based phylogenetic analysis) is introduced. The Qphyl system uses a new interactive approach to enable biologists to greatly improve the final results through effectively dynamic interaction with the computation, e.g., to move the computation back and forth to different stages so users can check the intermediate results, compare results from different methods and carry out certain manual refinements using their biological domain-specific knowledge in the decision making on how a tree should be reconstructed. Currently the alpha version of this web-based interactive system has been released and accessible through the URL: http://ww-test.it.usyd.edu.au/sogrid/qphyl/.